Performance analysis for, and interpretation of, data from MASB sonar in the application of swath bathymetry

Multi-Angle Swath Bathymetry (MASB) sonar produces high-quality, fully-registered bathymetry and imagery. This thesis presents a theoretical analysis of MASB sonar performance and example survey data, thereby providing a foundation for quantitative results, and demonstrating its potential as an effective survey tool. The quality of bathymetric surveys depends on the ability of a sonar to estimate the location of the bottom, therefore it is necessary to identify mechanisms that limit bottom estimation accuracy. Signal cross-correlations between elements on a linear array are determined for several transmitted pulse shapes for a general survey geometry. Decorrelation caused by geometrical mechanisms of footprint shift and baseline decorrelation are identified in the correlation functions. Utilizing signal correlations, the square root of the Cramer-Rao Lower bound (CRLB) on angle of arrival variance is identified as a better performance measure than the standard deviation. It is shown that for short pulse lengths, the dominant factors influencing performance are the footprint shift effect, the signal to noise ratio, and the presence of multiple signals on the array. A new measure of performance, error arc length (EAL), is defined using the CRLB, and EALs are plotted for a range of survey conditions including various frequencies, array tilt angles and number of array elements for both salt and fresh water surveys. To demonstrate the capabilities of MASB sonar, a prototype system was designed, constructed, and deployed. Pavilion Lake was chosen as one site for experimentation, as it is the focus of an international astrobiological effort to examine microbialites. Using MASB surveys, deep water microbialites were discovered possessing a morphology unlike others examined in prior research at the lake. Due to difficulties measuring ground truth information and operating on a moving survey platform, bottom estimation accuracy could not be directly compared with theoretical EALs for measurements at Pavilion Lake. Measurements taken at Sasamat Lake are presented which demonstrate that, in principle, the EAL can be used to predict sonar bottom estimation performance. In conclusion, the theoretical foundation presented, and the survey imagery and co-located bathymetry produced with the prototype sonar, demonstrate the effectiveness of MASB sonar for shallow water survey applications.